polymers

Article
Modification of Rule of Mixtures for Tensile Strength
Estimation of Circular GFRP Rebars
Young-Jun You 1, *, Jang-Ho Jay Kim 2 , Ki-Tae Park 1 , Dong-Woo Seo 1 and Tae-Hee Lee 2
1 Structural Engineering Research Institute, Korea Institute of Civil Engineering and Building Technology,
283 Goyangdae-Ro, Ilsanseo-Gu, Goyang-Si 10223, Korea; ktpark@kict.re.kr (K.-T.P.);
dwseo@kict.re.kr (D.-W.S.)
2 Concrete Structural Engineering Laboratory, School of Civil and Environmental Engineering,
Yonsei University, 50 Yonsei-Ro, Seodaemun-Gu, Seoul 03722, Korea; jjhkim@yonsei.ac.kr (J.-H.J.K.);
saintlth@yonsei.ac.kr (T.-H.L.)
* Correspondence: yjyou@kict.re.kr; Tel.: +82-31-910-0052

Received: 7 November 2017; Accepted: 4 December 2017; Published: 7 December 2017

Abstract: The rule of mixtures (ROM) method is often used to estimate the tensile strength of fiber
reinforced polymers (FRPs) reinforcing bars (rebars). Generally, the ROM method predicts the FRP
rebars’ modulus of elasticity adequately but overestimates their tensile strength. This may result
from defects occurred during manufacture that prevent the used materials from exhibiting a sound
performance and the shear-lag phenomenon by transmission of external forces through the surface
of the rebar having a circular cross section. Due to the latter, there is a difference in fiber breaking
points regarding the fibers located on the surface and fibers located at the center, and thus results in
differences between the values calculated from the conventional ROM and the experimental result.
In this study, for the purpose of resolving the problem, glass FRP (GFRP) rebars were shaped to
have a hollow section at the center of their cross sections and were further subject to tensile strength
tests. The test results were further placed under regression analysis and a modified ROM within
±5% accuracy compared to the experimental value was proposed for GFRP rebars with 13, 16, and 19
mm diameters.

Keywords: FRP; rebar; rule of mixtures; tensile strength; hollow section

1. Introduction
Fiber reinforced polymers (FRPs) have been mainly utilized in the aeronautical, aerospace,
and automotive fields owing to their relatively high strength, lightweight, and non-corroding qualities.
The construction sector began paying attention to FRPs in the 1950s as they were recognized as a
replacing material capable of solving the problem of degradation of structural performance caused
by the corrosion of steel reinforcement in concrete structures. Accordingly, efforts were made to
develop FRP reinforcing bars (rebars) in the 1960s and these efforts resulted in the major utilization of
FRPs in structures in the 1970s [1]. Internationally, active research on FRP rebars in the U.S., Canada,
and Europe led to the commercialization and application of several products in structure construction.
For economic reasons, glass fiber is often used rather than carbon or aramid fiber in the production
of FRP rebars [2]. Even if the tensile strength of bars made with glass fiber reinforced polymers (GFRPs)
attains strengths of typically around 700 MPa (ISIS 2007), recent achievements have succeeded in
developing GFRP rebars with strengths higher than 1000 MPa [3,4].
If the FRP rebars were manufactured perfectly without defects and uniform load applies within
the cross section, the materials used to manufacture the FRP rebars are expected to exhibit a sound
performance, in which the tested tensile strength values are the same as the value calculated from
previous ROM. However, due to production defects that occur as a result of the tangling of fibers

Polymers 2017, 9, 682; doi:10.3390/polym9120682 www.mdpi.com/journal/polymers

Polymers 2017, 9, 682 2 of 13

or appearance of voids, the used materials may not exhibit a sufficiently sound performance [5].
In addition, in general, GFRP rebars intended to replace steel rebars have circular cross-sections with
deformed or sand-coated surfaces to properly achieve the transfer of forces to the rebar in concrete.
This causes the rebar to experience shear-lag where the fibers located in the periphery of the rebar
cross-section receive higher stress than those located at the center. Therefore, unlike the steel rebar,
the tensile strength of the FRP rebar is a function of the rebar’s diameter [6].
For this reason, the rule of mixtures (ROM) method, which evaluates the tensile strength of FRPs
using linear elastic material qualities, tends to overestimate the tensile strength of FRP rebars. It is
thus necessary to modify the conventional ROM method to estimate properly the tensile strength of
FRP rebars.
Accordingly, this study intends to derive a ROM equation for GFRP rebars assuming that the
shear-lag effect causes the stress across the rebar cross-section to be distributed along a quadratic
curve when the GFRP rebar is tensioned. To that end, adjustment coefficients within ±5% accuracy
compared to the experimental value are proposed based upon the tensile test results of GFRP rebars
with hollow sections and diameters of 13, 16, and 19 mm.
The tensile strength performance of GFRP rebars may differ according to the type of material
used (for example, E (electrical)-, S (strength)-, and AR (alkali resistant)-glass) and the method by
which they were manufactured. Today, several different types of GFRP rebars have been developed
and there exist structural design guidelines regarding the application of some products. However,
the development of GFRP rebars is currently a work in progress. In light of this, despite the results of
this study not being appropriate for application to all GFRP rebars, the GFRP rebars considered in this
research were used to propose a modified ROM that considers only one variable, which was focused
on the sound performance of the fiber used in the GFRP rebar (relatively higher tensile strength/fiber
content than previously researched GFRP rebars). Due to the linear material feature, this model may
help other researchers as an index in the design and evaluation of GRFP rebar through any method
like linear interpolation.

2. Background
The rule of mixtures (ROM) is a weighted mean used to predict the properties of composite
materials such as FRPs including the tensile performance based upon the following assumptions [7]:

(1) One ply is microscopically homogenous, linear elastic, and orthotropic. In addition, it is initially
in a stress-free state.
(2) The fiber is homogenous, linear elastic, and well-arranged regularly in space.
(3) The matrix is also homogenous, linear elastic, and isotropic.
(4) There are no voids, and the fiber and the matrix are completely coupled.

Based on these assumptions, the tensile performance of FRP composed of fiber and polymer
matrix can be obtained by combining linearly the volume fraction and the tensile properties of the
fiber and the matrix as follows [8]:
σFRP = σ f Vf + σm Vm (1)

EFRP = E f Vf + Em Vm (2)

where σFRP , σ f , and σm , and EFRP , E f , and Em indicate the tensile strength (MPa) and the elastic
modulus (MPa) of FRP, reinforcing fiber, and matrix, respectively. Vm and Vf are the volume fractions
(%) of the reinforcing fiber and matrix, respectively.
ROM is the simplest and easiest method to predict the tensile properties of FRP with a given fiber
volume fraction and using the characteristics of its components. However, although its prediction of
the tensile elastic modulus in the axial direction is effective or accurate, ROM fails to predict accurately
the tensile strength [9]. In addition, ROM assumes that the fiber is unidirectionally aligned and
stress is uniformly distributed. In reality, the spread of the fiber can be non-homogenous and the
Polymers 2017,
Polymers 9, 682
2017, 9, 682 33of
of13
13
Polymers 2017, 9, 682 3 of 13

aligned and stress is uniformly distributed. In reality, the spread of the fiber can be non-homogenous
aligned
fiber and stress
orientation is be
can uniformly distributed.
misaligned, In in
resulting reality, the spread
reduced of the fiber
tensiletensile
strength of can be
theof non-homogenous
unidirectional fiber
and the fiber orientation can be misaligned, resulting in reduced strength the unidirectional
and the fiber
composite orientation can be misaligned, resulting in reduced tensile strength of the unidirectional
[10].
fiber composite [10].
fiberMoreover,
compositeROM [10]. also considers that the fiber and the matrix experience identical deformation,
Moreover, ROM also considers that the fiber and the matrix experience identical deformation,
Moreover,
as shown
shown ROM
in Figure
Figure 1. also considers
However, sincethat thematerials
both fiber andexhibit
the matrix experience
different tensileidentical deformation,
behaviors, shear-lag
as in 1. However, since both materials exhibit different tensile behaviors, shear-lag
as shown
occurs and in Figure
causes 1. However,
rupture to occursince
at both materials
different points of exhibit
time. different tensile behaviors, shear-lag
occurs and causes rupture to occur at different points of time.
occurs and causes rupture to occur at different points of time.

Figure 1. Schematic illustration of deformation of FRP under stress [9].

Figure1.1.Schematic
Figure Schematicillustration
illustrationof
ofdeformation
deformationof
ofFRP
FRPunder
understress
stress[9].
[9].
When a rebar is tensioned by grips, shear-lag will not occur when the bar is made of a highly
Whenaarebar
When rebarisistensioned
tensioned
byby grips,
grips, shear-lag
shear-lag willwill
not not
occuroccur when bar
the isbar is made of a highly
stiff material such as steel. However, as shown in Figure 2, this is when
not thethe
case for made
FRP of a highly
which stiff
combines
stiff material
material such such
as as steel.
steel. However,
However, as as shown
shown in in Figure
Figure 2, 2, this
this is is not
not the
the casefor
case forFRP
FRPwhich
whichcombines
combines
materials with relatively high stiffness (fiber) and relatively low stiffness (resin). This explains why
materialswith
materials withrelatively
relativelyhigh
highstiffness
stiffness(fiber)
(fiber)and
andrelatively
relativelylowlowstiffness
stiffness(resin).
(resin).This
Thisexplains
explainswhy
why
ROM predicts the elastic modulus of FRP rebars with relatively good accuracy but overestimates the
ROMpredicts
ROM predictsthe
the elastic modulus of FRP rebars with relatively good accuracy but overestimates the
tensile strength. elastic modulus of FRP rebars with relatively good accuracy but overestimates the
tensilestrength.
tensile strength.

Figure 2. Shear-lag phenomenon: (a) overview; and (b) longitudinal section.

Figure2.2.Shear-lag
Figure Shear-lagphenomenon:
phenomenon:(a)
(a)overview;
overview;and
and(b)
(b)longitudinal
longitudinalsection.
section.
Accordingly, Lee and Hwang [10] proposed the modified ROM expressed in Equations (3) and
Accordingly, Lee Lee and Hwang [10] proposed thethemodified ROM expressed in Equations (3) and
(4), inAccordingly,
which an effectiveand fiber Hwang
volume[10] proposed
fraction is applied modified
by meansROM expressed
of a degradation in Equations
parameter. The(3)
(4), in which
and (4), inare whichan effective
an effective fiber volume
fiber volume fraction
fraction is applied by means
is applied by means of a degradation
of avariation
degradation parameter.
parameter.The
equations based on the assumption that the ROM-originated strength is affected by
equations
The equations are arebasedbased onon thetheassumption
assumption that
that the
the ROM-originated
ROM-originated strength
strength variation
variation is
is affected by
affected by
factors such as the non-homogenous fiber spread in the case of a small fiber volume fraction and the
factors
factors such as the non-homogenous fiber spread in the case of a small fiber volume fraction and the
lack of asuch as the
matrix non-homogenous
between some adjacent fiber spread in the case of a small fiber volume fraction and the
fibers.
lackof
lack ofaamatrix
matrixbetween
betweensome someadjacent
adjacentfibers.fibers.
 FRP   f V fe   m (1  V f ) (3)
 FRP=σ Vf V fe+σ (m1(1−VV)f )
σFRP (3)
(3)
f fe m f
V fe  V f (1  P) (4)
VVf efe=V V f((11− PP))
f
(4)
(4)
where V fe is the effective volume fraction of reinforcing fiber (%) and P is a degradation
whereVfVe feis the
where effective
is the effectivevolume volume fraction of reinforcing
fraction of reinforcing fiber (%) and(%)
fiber P isand
a degradation parameter.
P is a degradation
parameter.
Other symbolsOther symbols
are symbols
same as thoseare same as those
in Equations in Equations
(1)Equations (1) and
and (2). (1) and (2). (2).
parameter. Other are same as those in
Faza
Faza and Gangarao [11] developed an analytical model to estimate the tensile strength of FRP
Faza and Gangarao [11] developed an analyticalmodel
and Gangarao [11] developed an analytical model to toestimate
estimatethe thetensile
tensilestrength
strengthof ofFRP
FRP
rebars
rebars based
based on
on the
the mechanics
mechanics of
of materials.
materials. This
This model
model considered
considered the
the shear-lag
shear-lag effect
effect by
by assuming
assuming
rebars based on the mechanics of materials. This model considered the shear-lag effect by assuming
that
that the strain of the bar resulting from the tension introduced by the gripping system is parabolic
thatthe
thestrain
strainof ofthethebarbarresulting
resultingfrom fromthe thetension
tensionintroduced
introducedby bythe thegripping
grippingsystem
systemisisparabolic
parabolic
and
and axisymmetric in the circular cross-section as shown in Figure 3. The parameter c denoting
parameter cc denoting the
andaxisymmetric
axisymmetric in in the
the circular
circular cross-section
cross-section as as shown
shown in in Figure
Figure 3. 3. The
The parameter denotingthe the
thickness
thickness of
of the
the boundary
boundary layer
layer was
was introduced
introduced to
to relate
relate the
the curing
curing rates
rates associated
associated with
with the
the rebar’s
rebar’s
thickness of the boundary layer was introduced to relate the curing rates associated with the rebar’s
size.
size. To establish their model expressed in Equation (5), the authors assumed the strain distribution
size.ToToestablish
establishtheir theirmodel
modelexpressed
expressedin inEquation
Equation(5), (5),the
theauthors
authorsassumedassumedthe thestrain
straindistribution
distribution
shown
shown in
inFigure
Figure 3.3.InInthis
this distribution,
distribution, the
thestrain
strain diminishes
diminishes parabolically
parabolically from
fromthethe
outermost
outermost part of
part
shown in Figure 3. In this distribution, the strain diminishes parabolically from the outermost part of
the rebar to the height of D / 2  c−, where D is the diameter of the FRP rebar, and shows even
the rebar to the height of D / 2  c , where D is the diameter of the FRP rebar, and shows even
of the rebar to the height of D/2 c, where D is the diameter of the FRP rebar, and shows even
distribution
distributionafterward
afterwardto tothe
thecenter
centerwith withthe thefiber’s
fiber’sextremeextremestrain decreased bybyλ..
distribution afterward to the center with the fiber’s extreme strain decreased by  .
strain decreased

σFRP 0.5f [f1[+

=0.5σ 1 2λ 2−((1 −22c/Dc / D)2)22]] (5)(5)
FRP
FRP  0 .5  f [ 1  2  (1  2c / D) ] (5)
Polymers 2017, 9, 682 4 of 13
Polymers 2017, 9, 682 4 of 13

Figure 3. Strain distribution across cross-section of FRP rebar in tension [11].

Equation (5)
Equation (5)predicts
predictsthe
the actual
actual tensile
tensile strength
strength of FRP
of the the FRP
rebarsrebars
quite quite well adequate
well when when adequate
values
values
are used are
forused for the parameters.
the parameters. However, However, the determination
the determination of parameters
of the two the two parameters c and on
c and λ depends 
depends
the curingonrate
theofcuring rateand
the resin of the
theresin and
size of the
the size of therespectively.
specimens, specimens, respectively.
Despite it being practical to consider performance losses due due
Despite it being practical to consider performance losses to incomplete
to incomplete bondsbonds
between between
fibers,
fibers,within
voids voids the
within the composite,
composite, and the non-homogeneous
and the non-homogeneous spread
spread of fibers, of elements
such fibers, such elements of
of incompletion
incompletion
for for casessmall
cases of relatively of relatively small
bar shaped bar shaped
diameters anddiameters and long
long lengths may lengths
have verymay have
little veryor
effect, little
an
effect, or an extremely large effort is needed to quantify all such elements
extremely large effort is needed to quantify all such elements of incompletion. of incompletion.

3. Model for Tensile Strength Estimation

Tensile Strength Estimation
The tensile strength of the FRP rebars is sensitive not only to the amount and material properties
of the fibers but also to to various
various other conditions such as the manufacturing process and fiber
arrangement [5]. Moreover, since reinforced
reinforced fibers
fibers are
are highly
highly vulnerable
vulnerable to lateral loading, the tensile
according to the grip system used to apply tension as well
strength of the FRP rebars may differ even according
between the
as to the bond performance between the grip
grip and
and the
the FRP
FRP rebar
rebar surface
surface [12,13].
[12,13].
Considering the difficulty of taking into account all such manufacturing
Considering the difficulty of taking into account all such manufacturing conditions,
conditions, the
the following
following preconditions
preconditions were assumed weremacroscopically
assumed macroscopically
in this studyinto this studyantoequation
develop developpredicting
an equation
the
predicting
tensile the tensile
strength of FRPstrength
rebars. of FRP rebars.
(1) FRP
(1) FRP rebars
rebars are
are made
made only
only of
of aa single
single type
type ofof reinforced fiber (especially
reinforced fiber (especially glass
glass fiber)
fiber) and
and resin.
resin.
(2) FRP
(2) FRP rebars
rebars are
are fabricated
fabricated through
through an an adequate
adequate process,
process, and
and the
the so-produced
so-produced rebars exhibit small
rebars exhibit small
variation regarding the tensile
variation regarding the tensile strength. strength.
(3) Because
(3) Because the the values
values of
of material
material properties
properties of of resin
resin are
are far
far smaller
smaller than
than those
those ofof reinforcing
reinforcing fiber,
fiber,
itit is assumed that, for the sake of model simplification, the tensile strength
is assumed that, for the sake of model simplification, the tensile strength of the FRP rebars is of the FRP rebars is
controlled macroscopically only by the quantity of the
controlled macroscopically only by the quantity of the reinforcing fiber. reinforcing fiber.
(4) There is no transverse shrinkage perpendicular to the longitudinal axis of the rebar.
(4) There is no transverse shrinkage perpendicular to the longitudinal axis of the rebar.
(5) The grip used to evaluate the tensile strength of the FRP rebars supports the rebars sufficiently
(5) The grip used to evaluate the tensile strength of the FRP rebars supports the rebars sufficiently
up to their breaking load or displays the same performance.
up to their breaking load or displays the same performance.
(6) The conventional ROM predicts accurately enough the elastic modulus of FRP rebars.
(6) The conventional ROM predicts accurately enough the elastic modulus of FRP rebars.
The equation predicting the tensile strength of the FRP rebars proposed in this research
resemblesequation
The predicting
that suggested by the
Fazatensile
and strength
Gangarao of[11].
the FRP rebars proposed
However, considering in this
thatresearch resembles
the diameter and
that suggested by Faza and Gangarao [11]. However, considering that the
length of the rebars are sufficiently comparable, the equation is formulated by ignoring the sectiondiameter and length of the
rebars
where are thesufficiently comparable,
strain is distributed the equation
evenly is formulated
and by assuming that bytheignoring the section
distribution where
of stress the inside
occurs strain
is
thedistributed
cross-section evenly and
as the by assuming
quadratic curvethat the distribution
in Equation (6). of stress occurs inside the cross-section as
the quadratic curve in Equation (6).
If the distribution of the stress developed in the FRP rebars under tension and close to rupture
If the distribution
is axisymmetric and takesof the
the stress
form of developed
a quadratic in curve,
the FRP therebars
stressunder tensionacross
distribution and close to rupture
the rebar cross-
is axisymmetric and takes the form
section can be expressed as shown in Figure 4. of a quadratic curve, the stress distribution across the rebar
cross-section can be expressed as shown in Figure 4.
f ( x)  Ax 2  Bx  C (6)
f ( x ) = Ax2 + Bx + C (6)
Polymers 2017, 9, 682 5 of 13
Polymers 2017, 9, 682 5 of 13
Polymers 2017, 9, 682 5 of 13

Figure 4. Assumed stress distribution on transversally projected section.

Figure
Figure 4.
4. Assumed
Assumed stress
stress distribution
distribution on
on transversally projected section.
In Figure 4,  fu is the ultimate tensile strength of FRP depending on the fiber volume fraction
4, σf ufu isisthe
f and D are the tensile strength and the diameter of a rebar, respectively;  is a radius
( V f );InFigure
In Figure 4, theultimate
ultimatetensile
tensilestrength
strengthofofFRP
FRPdepending
depending on on the
the fiber
fiber volume
volume fraction
fraction
(V
f ); 
);
( Vfany
of
σ f and
f and
point
D are
from
the
Dthe tensile
arecenter strength
the tensile
(which
and
strength the
means and
diameter of
the diameter
the radius
a rebar, respectively;
of a section
of hollow
δ
rebar, respectively;is
in this paper);
a  and
radius of any
is a radius
is a
point from the center (which means the radius of hollow section in this paper); and γ is a reduction
means the radius of hollow section in this paper); and  is a
< γ < 1).(0 <  < 1).
of any point
reduction from
factor of the
the center
tensile(0(which
strength
factor of the tensile strength
reduction
The
The sizefactor
size of of the
ofthe
the tensile strength
distributed
distributed stress
stress (0be
can
can  obtained
< be < 1). by by
obtained subtracting
subtracting the volume
the volume of theofparaboloid
the paraboloidbowl
bowl from
The the
size ofcircular
the cylinder
distributed in Figure
stress can 5.
be The volume
obtained by is computed
subtracting theas the
volume
from the circular cylinder in Figure 5. The volume is computed as the height of the stress multiplied height
of theof the stress
paraboloid
multiplied
bowl
by thefrom areaby thethe
and area and
circular
means means
cylinder
thus thus
in
the force. the force.
Figure 5. The volume is computed as the height of the stress
multiplied by the area and means thus the force.

Figure 5.
Figure Calculation of
5. Calculation of peak
peak tensile
tensile force.
force.
Figure 5. Calculation of peak tensile force.
In this
this stress
stressdistribution,
distribution,the ultimate
the ultimate tensile strength
tensile strength(σ f u() 
occurs acrossacross
fu ) occurs the rebar’s outer surface
the rebar’s outer
surface while the ultimate tensile strength reduced by  is at work
while Inthe ultimate
this stress tensile strength
distribution, reduced
the ultimate by is at
tensile
γ work in
strength the
( center
) of
occurs the cross-section.
across the Therefore,
rebar’s
fu in the center of the cross-section. outer

side2 passes
the quadratic curve projected onto the side passes through (0, γσ ), (D/2, σ ), and ( − D/2, σ f u ),
surface while
Therefore, thethe ultimatecurve
quadratic tensileprojected
strength reduced
onto the by is at work f in
through (0,  fuof), the
u the center fu
 fu ), and
cross-section.
(D/2,
leading to C = γσ f u , B = 0 and A = (1 − γ)σ f u /( D/2) .
Therefore, the quadratic curve projected onto the side passes through (0,  fu ), (D/2,  fu ), and
(−D/2,In  fu ), leading to C   fu , B  0 and A  (1   ) fu /( D / 2) .
Figures 4 and 5, the volume F1 of the cylinder with radius δ and 2 height f (δ), and the volume
(−D/2,  fu ), leading to C   fu , B  0 and A  (1   ) fu /( D / 2) .
F2 of the paraboloid bowl can be obtained as follows: 2
In Figures 4 and 5, the volume F1 of the cylinder with radius ) and height f ( ) , and the
 and height f ( ) , and the
(
In Figures 4 and
volume F2 of the paraboloid 5, the volume
bowl F of the cylinder
( 1 − γ )with
σ f u radius
F1 = πδ2can
f (1δ)be=obtained
πδ2 as follows:
2
δ2 + γσ f u (7)
volume F2 of the paraboloid bowl can be obtained(as D/2follows:
)
 (1   ) 
F1   2 f (Z )f (δ)  2  (1  (1)−2fuγ)σ2   fu  (7)
2 2  (D 
F1  F    
π / 2) fu 2f u 4
2
2
= f
π ( )  x dy  2
= 2 2 δ fu  (8)
(7)
γσ f u
 ( D / 2( )D/2) 
f ( )  (1   )
F2    f ( ) Fx 2= dyF1− F2(1   )2fu  4 (9)
(8)
 fu 2 ( D / 2) fu 4
 fu xδdyand hollow
2  radius 
2
Considering the FRP rebarFwith 2 ( D /part
(8)
2) 2 in the cross, as shown in Figure 5,
the volume F Empty of the removed part of the F composite
F1  F2 volume can be obtained by Equation(9) (9),
F  F1  F2 (9)
Considering the FRP rebar with radius  and hollow part in the cross, as shown in Figure 5,
the volume F EmptytheofFRP
Considering therebar withpart
removed radius  composite
of the and hollow part incan
volume thebe
cross, as shown
obtained in Figure(9),
by Equation 5,
the volume F Empty of the removed part of the composite volume can be obtained by Equation (9),
Polymers 2017, 9, 682 6 of 13

and the volume of the full cross-section F Full can also be obtained by Equation (9) in which δ in
Equations (7) and (8) is replaced by D/2. That is,
( )
Empty Empty Empty π (1 − γ ) σ f u
F = F1 − F2 = δ2 σ f u 2
δ + 2γ (10)
2 ( D/2)2
 2    
Full D 1+γ 1+γ
F =π σf u = A FRP Vf σ f u (11)
2 2 2
where A FRP is the cross-sectional area of the FRP rebar.
Consequently, Equation (11) representing the tensile breaking force of the rebar is a modified
ROM because it multiplies Equation (1) by the cross-section area of the rebar and applies the tensile
strength reduction rate.
Assuming that, with the exception of the reduced tensile strength, the tensile behavior of the FRP
rebar with a hollow section is not different from that of the FRP rebar with a full cross section in terms
of the elastic modulus, the tensile strength of the hollow section FRP rebar can be obtained as follows.
( )
(1 − γ ) σ f u 2
 
Hollow Full Empty 1+γ π 2
F =F −F = A FRP Vf σ f u − δ σ f u δ + 2γ (12)
2 2 ( D/2)2

The tensile strength of the FRP rebar can be obtained adequately, if the tensile strength change
rate γ can be obtained from Equation (12) through tension test of the hollow FRP rebar with parts
missing in the center of the cross-section.
Unlike the previous ROM model that assumes the external force to be equally distributed across
the FRP cross-section, this model features a bar shaped FRP having a circular cross-section that
considers actual situations in which external forces are not transmitted to the cross-section but from
the side section surface. Due to there being only one variable, it is easier to apply than Equation (5)
that uses two variables.

4. Tensile Test of Hollow GFRP Rebar Specimens

4.1. Materials and Manufacturing Process

The GFRP rebars used in this study are made of E-glass fiber (Owens Corning Korea, Seoul, Korea)
and vinyl ester resin (Aekyung Chemical Co., Ltd., Seoul, Korea, Ashland Inc., Covington, KY, UAS).
The properties of these materials by manufacturers are listed in Table 1.

Table 1. Material properties of GFRP (glass fiber reinforced polymers) rebar.

Material Designation of Product Tensile Strength (MPa) Elastic Modulus (MPa)

Resin 1 HETRON 922 86 3170
Resin 2 DION-9100 79 3216
Core fiber SE1200 2600 81,000

The rebars were fabricated by the braidtrusion process of You et al. [5], which is a modified
version of the one suggested by Ko et al. [14]. This modified process enhances the overall tensile
performance through improved fiber arrangement and reduced voids by applying a definite tension to
the fibers used to form reinforced fiber bundles and ribs.

4.2. Hollow GFRP Rebar Specimens

Sets of GRFP rebars with two different diameters and four different hollowness ratios were
fabricated to investigate the changes in the tensile properties according to the hollowness ratio of
the GFRP rebars. Plain GFRP rebars (in other words, rebars with solid cross-section) were also
Polymers
Polymers 2017,
2017, 9,
9, 682
682 77 of
of 13
13
Polymers 2017, 9, 682 7 of 13
to
to provide
provide reference
reference specimens
specimens forfor the
the comparison
comparison of of the
the tensile
tensile properties
properties with
with respect
respect to
to the
the
hollowness ratio. The hollow GFRP rebars were manufactured by inserting tubes
hollowness ratio. The hollow GFRP rebars were manufactured by inserting tubes with different with different
considered to provide reference specimens for the comparison of the tensile properties with respect to
diameters
diameters inside
inside the
the rebars.
rebars. Table
Table 22 lists
lists the
the considered
considered specimens
specimens with
with their
their designation,
designation,
the hollowness ratio. The hollow GFRP rebars were manufactured by inserting tubes with different
specifications,
specifications, and
and quantities.
quantities.
diameters inside the rebars. Table 2 lists the considered specimens with their designation, specifications,
and quantities. Table 2. Test specimens.
Table 2. Test specimens.
Specimen
Specimen Quantity (mm)2. Test
DiameterTable specimens.
Quantity Diameter (mm) Ratio of Hollow Section Number
Ratio of Hollow Section
Number of of Roving
Roving
D16HD0
D16HD0 55
Specimen Quantity 15.9 -- Section Number of158
15.9 (mm) Ratio of Hollow
Diameter 158
Roving
D16HD6
D16HD6
D16HD0
66 5 16.0
16.0
15.9
14.1%
14.1%
-
129
158129
D16HD8
D16HD8
D16HD6 66 6 16.0
16.0
16.0 25.0%
25.0%
14.1% 129112
112
D16HD10
D16HD8
D16HD10 66 6 16.0
16.0
16.0 39.1%
25.0%
39.1% 1129191
D16HD10
D16HD12 6 6 16.0
16.0 39.1%
56.3% 91 65
D16HD12 6 16.0 56.3% 65
D16HD12 6 16.0 56.3% 65
D19HD0
D19HD0 33 18.6
18.6 -- 217
217
D19HD0 3 18.6 - 217
D19HD6
D19HD6
D19HD6 66 6 18.6
18.6
18.6 10.4%
10.4%
10.4% 196196
196
D19HD8
D19HD8
D19HD8 66 6 18.7
18.7
18.7 18.3%
18.3%
18.3% 180180
180
D19HD10
D19HD10 66 6 18.8
18.8 28.4%
28.4% 159159
D19HD10 18.8 28.4% 159
D19HD12
D19HD12 6 6 19.0
19.0 39.7%
39.7% 133133
D19HD12 6 19.0 39.7% 133

In
In the
the specimen designation of
specimen designation
designation of Table
Table 2,
2, D D followed
followed by by aa two-digit
two-digit number
number denotes
denotes the
the outer
outer
diameter of the GFRP rebar, and HD indicates the diameter of the hollow tube
diameter of the GFRP rebar, and HD indicates the diameter of the hollow tube inserted inside the
of the GFRP rebar, and HD indicates the diameter of the hollow inserted
tube inside
inserted the GFRP
inside the
rebar.
GFRP For
GFRP rebar. instance,
rebar. For D19HD8
For instance,
instance, designates
D19HD8
D19HD8 the specimen
designates
designates the with awith
the specimen
specimen 19-mm
with diameter
aa 19-mm
19-mm for thefor
diameter
diameter whole
for the GFRP
the whole
whole
rebar
GFRPcross-section
GFRP rebar in which
rebar cross-section
cross-section inawhich
in hollowaa tube
which hollow
hollowwith an with
tube
tube external
with an diameterdiameter
an external
external of 8 mm is
diameter ofinserted.
of 88 mm
mm is The diameter
is inserted.
inserted. The
The
of the reinforcing
diameter of the bars in
reinforcing Table 2
bars is
in measured
Table 2 is at parts
measured without
at parts
diameter of the reinforcing bars in Table 2 is measured at parts without ribs. ribs.
without ribs.
Polyurethane
Polyurethane tubestubes were used to
were used
used form the
to form
form the hollow
the hollow section
hollow section within
section within the
within the GFRP
the GFRP rebar
GFRP rebar cross
rebar cross section.
cross section.
section.
As
As shown
shown in in Figure
Figure 6,6, aa fixture
fixture was
was used
used to
to control
control the
the centrality
centrality ofof the
the tube
tube inside
inside the
the glass
glass fiber
fiber
bundles. Figure 7 shows the cross-sections of the so-fabricated hollow
bundles. Figure 7 shows the cross-sections of the so-fabricated hollow GFRP rebars. GFRP rebars.
rebars.

Figure 6.
Figure Tube insertion
6. Tube
Tube insertion for
for hollow
hollow section.
section.

Figure
Figure 7.
7. Cross-section
Cross-section of
of hollow
hollow GFRP
GFRP rebars (D16).
rebars (D16).
 The finished GFRP rebars were cut to a definite length, and cylindrical steel pipes were
positioned at the ends to form the grip in compliance with CSA [15]. These steel pipes have respective
thickness and length of 5.1 mm and 700 mm for the D16 rebars, and 7.1 mm and 1000 mm for the D19
rebars. The caps placed on both ends of the steel pipes are hollowed at their center to allow the rebar
Polymers 2017, 9, 682 8 of 13
be inserted
Polymers centrally
2017, 9, 682 in the steel pipes. The grips were first made on one side of the rebar and the
8 of 13
space between the steel pipe and the rebar was filled with non-shrinkage mortar. The rebars were
placed finished
Thestraight
finished GFRP
byGFRP
means rebars
rebars werewere
of holders cut
cutand
to to a definite
acuring
definitewas length,
conducted
length, andseven
for
and cylindrical cylindrical
days.
steel pipes steel
werepipes
Figure were
8 shows
positioned a
positioned
sample
at at
to the
of GFRP
the ends form ends
theto
rebar formin the
specimens
grip grip in compliance
prepared
compliance for tensile
with with
CSA [15]. CSA [15].
test.These steelThese
pipessteel
havepipes have respective
respective thickness
thickness
and lengthand length
of 5.1 mmofand5.1 700
mmmm and for
700the
mmD16 for rebars,
the D16and rebars, and and
7.1 mm 7.1 mm1000and
mm 1000
formm for the
the D19 D19
rebars.
rebars.
The The
caps caps placed
placed on both onends
bothofends
the of the pipes
steel steel pipes are hollowed
are hollowed at their
at their center
center to allow
to allow the the rebar
rebar be
be inserted
inserted centrally
centrally in the
in the steelsteel pipes.
pipes. The The
gripsgrips
werewere first made
first made on one onside
oneof side
theof the and
rebar rebartheand the
space
Nut
space between
between the steel thepipe
steeland
pipetheand thewas
rebar rebar waswith
filled filled with non-shrinkage
non-shrinkage mortar.mortar. The were
The rebars rebars were
placed
Screw
placed straight
straight by means byofmeans
holders of and
holders andwas
curing curing was conducted
conducted for seven for seven
days. days.
Figure Figurea 8sample
8 shows showsofa
sample of GFRP rebar specimens prepared
GFRP rebar specimens prepared for tensile test. for tensile test.

Steel tube

Nut
Mortar CapScrew Rebar

Figure 8. Preparation of specimens for tensile test.

4.3. Test Set-Up Steel tube

As shown in Figures 8 and 9, one end of the cylindrical steel pipe forming the grip was threaded
to fasten a nut to hold the specimen and apply tension
Mortar Cap on the Rebar
two ends of the grip.
The so-prepared specimen was then installed in a universal testing machine (UTM) with
Figure
Figure 8. Preparation of
8. Preparation of specimens
specimens for tensile test.
capacity of 1000 kN and an electrical resistance strain gauge for
wastensile test. to the middle of the rebar’s
attached
length to measure the strain during test. The place where the strain gauge was attached was cleaned
4.3. Test
Test Set-Up
Set-Up
softly, not grinded because the loss of cross-section would affect the tensile capacity. Tensile test was
As shown
performed in Figures
through 8 and 9, one
displacement end ofand
control the acylindrical steel(TDS530
data logger pipe forming the grip
produced in was threaded
Tokyo Sokki
to fasten a Co.,
Kenkyujo nut to hold
Ltd, the
the specimen
Tokyo, specimen
Japan) was and
and apply
apply
used tension on
tensionthe
to record the
the two
onload andends
two ends of
of the
the grip.
corresponding grip.strain.
The so-prepared specimen was then installed in a universal testing machine (UTM) with
capacity of 1000 kN and an electrical resistance strain gauge was attached to the middle of the rebar’s
length to measure the strain during test. The place where the strain gauge was attached was cleaned
softly, not grinded because the loss of cross-section would affect the tensile capacity. Tensile test was
performed through displacement control and a data logger (TDS530 produced in Tokyo Sokki
Kenkyujo Co., Ltd, Tokyo, Japan) was used to record the load and corresponding strain.

Figure 9. Tensile test set-up.

The so-prepared specimen was then installed in a universal testing machine (UTM) with capacity
5. Results
of 1000 kN and an electrical resistance strain gauge was attached to the middle of the rebar’s length to
5.1. Tensile
measure theBehavior of GFRP
strain during Rebar
test. The place where the strain gauge was attached was cleaned softly, not
grinded because the loss of cross-section would affect the tensile capacity. Tensile test was performed
Figure 10 plots the tensile stress–strain curves in which the stress is the UTM’s load divided by
through displacement control and a data logger
Figure (TDS530 produced in Tokyo Sokki Kenkyujo Co., Ltd.,
the area of specimen and the strain is the one9.measured
Tensile test
atset-up.
the middle of the rebar’s length. The curves
Tokyo, Japan) was used to record the load and corresponding strain.
indicate a linear relationship between the tensile strength and the strain curves, and the occurrence
5. Results
of sudden rupture at the peak load. This is the typical material behavior of FRP where the linear
5. Results
5.1. Tensile Behavior of GFRP Rebar
5.1. Tensile Behavior of GFRP Rebar
Figure 10 plots the tensile stress–strain curves in which the stress is the UTM’s load divided by
Figure 10 plots the tensile stress–strain curves in which the stress is the UTM’s load divided by
the area of specimen and the strain is the one measured at the middle of the rebar’s length. The curves
the area of specimen and the strain is the one measured at the middle of the rebar’s length. The curves
indicate a linear relationship between the tensile strength and the strain curves, and the occurrence
of sudden rupture at the peak load. This is the typical material behavior of FRP where the linear
Polymers 2017, 9, 682 9 of 13

Polymers 2017, 9, 682 9 of 13

indicate a linear relationship between the tensile strength and the strain curves, and the occurrence of
increase is followed
sudden rupture at thebypeak
brittle failure.
load. This The maximum
is the tensile behavior
typical material strength and thewhere
of FRP elasticthe
modulus change
linear increase
with respect by
is followed to the hollowness
brittle ratio,maximum
failure. The and tend to decrease
tensile with and
strength a larger
the hollowness ratio.change
elastic modulus Table 3 with
lists
the maximum tensile strength and elastic modulus of the considered specimens.
respect to the hollowness ratio, and tend to decrease with a larger hollowness ratio. Table 3 lists the The maximum
tensile
maximum strengths
tensilewere calculated
strength frommodulus
and elastic dividingofmeasured maximum
the considered loads by
specimens. Thesolid cross-section
maximum tensile
areas without considering the hollowness because the value calculated from
strengths were calculated from dividing measured maximum loads by solid cross-section areas without net cross-section area
with considering the hollowness means the tensile strength of a bar with smaller
considering the hollowness because the value calculated from net cross-section area with considering diameter and solid
section (for example,
the hollowness meansthe the net areastrength
tensile of a barofwith
a bar16 mmsmaller
with and 9.7 mm of and
diameter outer andsection
solid hollow(for
diameters,
example,
respectively, is same with the solid area of a bar with 12.7 mm diameter).
the net area of a bar with 16 mm and 9.7 mm of outer and hollow diameters, respectively, is same The measured average
with
diameters of the specimens were all 16.0 mm, except for D16HD0 (15.9 mm).
the solid area of a bar with 12.7 mm diameter). The measured average diameters of the specimens It appears that, apart
from
were specimens
all 16.0 mm, D16HD8 andD16HD0
except for D16HD12, themm).
(15.9 coefficient of variation
It appears (C.O.V)
that, apart fromobtained
specimens byD16HD8
dividing and
the
standard deviation (S.DEV) by the mean remains below 6.4% and indicates consistency
D16HD12, the coefficient of variation (C.O.V) obtained by dividing the standard deviation (S.DEV) by among the
experimental
the mean remains values.
below 6.4% and indicates consistency among the experimental values.

(a) (b)
Figure 10.
Figure Tensile behavior
10. Tensile behavior of
of specimens:
specimens: (a)
(a) D16;
D16; and
and (b)
(b) D19.
D19.

Table 3.
Table Tensile test
3. Tensile test results
results of
of plain
plain and hollow GFRP rebar specimens.

Peak Tensile Elastic Peak Tensile Elastic

Specimen
Specimen Tensile Elastic Specimen Tensile Elastic
Load Peak Strength Modulus Specimen Peak
Load Strength Modulus
Modulus
and
and Values Strength Modulus and Values Strength
Load (N) (MPa) and Values Load (N)
Values (N) (MPa) (MPa)
(MPa) (N) (MPa)
(MPa) (MPa)
(MPa)
D16HD0
D16HD0 213,191
213,191 10741074 49,402
49,402 D19HD0
D19HD0 330,210
330,210 1218
1218 53,520
53,520
209,767
209,767 10561056 46,798
46,798 333,300
333,300 1229
1229 55,180
55,180
213,848
213,848 10771077 50,681
50,681 306,760
306,760 1131
1131 56,519
56,519
meanmean 212,269 212,269 10691069 48,960
48,960 mean
mean 323,423
323,423 1193
1193 55,073
55,073
S.DEV.S.DEV. 2191 2191 11 11 1979
1979 S.DEV.
S.DEV. 14,513
14,513 5454 15021502
C.O.V 1.0% 1.0% 4.0% C.O.V 4.5% 4.5% 2.7%
C.O.V 1.0% 1.0% 4.0% C.O.V 4.5% 4.5% 2.7%
D16HD6 217,120
D16HD6 217,120 10801080 46,916
46,916 D19HD6
D19HD6 312,530
312,530 1147
1147 55,623
55,623
211,250
211,250 10511051 53,008
53,008 301,480
301,480 1106
1106 56,681
56,681
203,650 1013 46,237 308,260 1131 57,229
203,650 1013 46,237 308,260 1131 57,229
221,400 1101 51,907 280,370 1029 55,190
221,400
207,710 11011033 51,907
47,131 280,370
304,420 1029
1117 55,190
57,505
207,710
202,160 10331005 47,131- 304,420
277,080 1117
1016 57,505
59,276
mean 202,160
210,548 10051047 -
49,040 mean 277,080
297,357 1016
1091 59,276
56,917
meanS.DEV. 210,5487582 1047 38 3161
49,040 S.DEV.
mean 14,938
297,357 55
1091 146456,917
S.DEV.C.O.V 7582 3.6% 38 3.6% 6.4%
3161 C.O.V
S.DEV. 5.0%
14,938 5.0%
55 2.6%1464
D16HD8 3.6%
C.O.V 176,830 3.6%879 42,314
6.4% D19HD8
C.O.V 236,600
5.0% 863
5.0% 52,641
2.6%
D16HD8 178,360 879 887
176,830 41,436
42,314 D19HD8 246,140
236,600 898
863 49,577
52,641
176,520 878 40,710 248,740 908 50,267
178,360 887 41,436 246,140 898 49,577
179,780 894 49,002 230,930 843 54,674
176,520 878 40,710 248,740 908 50,267
179,780 894 49,002 230,930 843 54,674
181,870 905 41,146 239,260 873 50,036
172,470 858 - - - -
mean 177,638 884 42,922 mean 240,334 877 51,439
S.DEV. 3213 16 3449 S.DEV. 7209 26 2163
C.O.V 1.8% 1.8% 8.0% C.O.V 3.0% 3.0% 4.2%
Polymers 2017, 9, 682 10 of 13

Table 3. Cont.

Specimen Tensile Elastic Tensile Elastic

Peak Specimen Peak
and Strength Modulus Strength Modulus
Load (N) and Values Load (N)
Values (MPa) (MPa) (MPa) (MPa)
181,870 905 41,146 239,260 873 50,036
172,470 858 - - - -
mean 177,638 884 42,922 mean 240,334 877 51,439
S.DEV. 3213 16 3449 S.DEV. 7209 26 2163
C.O.V 1.8% 1.8% 8.0% C.O.V 3.0% 3.0% 4.2%
D16HD10 131,050 652 35,680 D19HD10 205,860 746 49,664
133,670 665 33,663 199,220 722 46,453
146,530 729 34,246 203,320 736 44,988
137,170 682 31,801 203,690 738 45,080
133,060 662 34,370 - - -
129,790 646 37,302 - - -
mean 135,212 672 34,510 mean 203,023 735 46,546
S.DEV. 6094 30 1861 S.DEV. 2772 10 2184
C.O.V 4.5% 4.5% 5.4% C.O.V 1.4% 1.4% 4.7%
D16HD12 86,330 429 28,130 D19HD12 194,500 684 38,819
89,500 445 29,762 207,820 731 35,442
97,260 484 26,079 182,040 640 34,792
92,350 459 26,601 207,730 730 33,410
83,030 413 22,041 189,390 666 37,175
90,580 451 25,959 189,740 667 32,999
mean 89,842 447 26,429 mean 195,203 686 35,439
S.DEV. 4914 24 2593 S.DEV. 10,521 37 2233
C.O.V 5.5% 5.5% 9.8% C.O.V 5.4% 5.4% 6.3%
S.DEV: standard deviation; C.O.V: coefficient of variation.

5.2. Regression Analysis

Regression analysis can be performed easily by the least squares method because the assumed
tensile stress distribution curve expressed in Equation (6) includes one independent variable and one
dependent variable. The resulting regression equation must be verified for each specimen to determine
its agreement with the measured values. This can be done using the coefficient of determination R2 of
which value runs between 0 and 1, with R2 = 1 indicating that the sample regression line fits perfectly
with the data.
The reduction factor γ of the tensile strength can be obtained by regression analysis of the values
in Table 3 and Equations (1) and (6)–(12). The results are γ = 0.31843 for D16 specimens and γ = 0.29549
for D19 specimens. Applying these values in Equation (11) leads to the following estimation of the
tensile strength in Equation (13) for D16 specimens and Equation (14) for D19 specimens.

Full
FD16 = 0.659A FRP Vf σ f (13)

Full
FD19 = 0.648A FRP Vf σ f (14)

As explained above, the tensile strength of the FRP rebars is a function of the rebar diameter
and tends to decrease with a larger diameter. Figure 11 indicates the change in the tensile strength
according to the diameter for the commercial product Aslan100 [16]. It appears that the tensile strength
of the reinforcing bar reduces quasi-linearly with the increase of the diameter.
Accordingly, assuming a linear relationship between the tensile strength and the diameter of the
GFRP reinforcing bar, the tensile strength can be estimated by Equation (15) for the 13 mm-diameter
rebar through linear interpolation of Equations (13) and (14).

Full
FD13 = 0.674A FRP Vf σ f (15)
 FD19  0.648 AFRPV f  f (14)

As explained above, the tensile strength of the FRP rebars is a function of the rebar diameter and
tends to decrease with a larger diameter. Figure 11 indicates the change in the tensile strength
according to the diameter for the commercial product Aslan100 [16]. It appears that the11tensile
Polymers 2017, 9, 682 of 13
strength of the reinforcing bar reduces quasi-linearly with the increase of the diameter.

Figure 11.11.
Figure Change inin
Change tensile
tensilestrength
strengthof
ofGFRP
GFRP rebar
rebar according todiameter
according to diameter[16].
[16].

5.3. Verification of Estimation Model

Although the actual fiber content is unknown because burn-out test was not performed on the
specimens considered in this research, many researchers reported that the ROM provides adequate
predictions of the elastic modulus of FRP materials. Thus, if the influence of the resin is ignored, as in
this study, the fiber content listed in Table 4 can be estimated by Equation (2) and by ignoring the
contribution of the matrix to the tensile strength.

Table 4. Calculation of fiber volume fraction.

Measured Inner Elastic Modulus (GPa)

Rebar Calculated Volume Fraction
Diameter (mm) Glass Fiber a Rebar b
D13 [5] 12.7 73 49 67.6%
D13 [17] 12.7 81 52 63.8%
D13 [18] 12.5 - 47 55.0%
D16 16.0 81 49 60.4%
D19 [19] 18.6 73 47 64.0%
D19 18.6 81 55 68.0%
a Provided by manufacturer; b Measured experimentally.

Table 5 compares the tensile strengths predicted by the conventional and modified ROMs to
validate the developed estimation model using the experimental results of this study and those of
previous tensile strength tests carried out by other researchers. It appears that the modified ROM
provides accuracy within ±5%, whereas the conventional ROM overestimates the tensile strength of
the GFRP reinforcing bars by more than 40%.

Table 5. Comparison of tensile strengths predicted by conventional ROM and modified ROM using
Equations (13)–(15).

Tensile Strength
Predicted Tensile Strength (MPa)
(MPa) Modification
Rebar
Glass Factor Modified Modified
Rebar b ROM * ROM/Experiment
Fiber a ROM ROM/Experiment
D13 [5] 2580 1132 0.674 1745 154% 1174 104%
D13 [17] 2600 1103 0.674 1660 150% 1117 101%
D13 [18] 2600 975 0.674 1430 147% 965 99%
D16 2600 1069 0.659 1572 147% 1036 97%
D19 [19] 2580 1030 0.648 1651 160% 1069 104%
D19 2600 1193 0.648 1768 148% 1145 96%
* From Equations (1) and (2).
Polymers 2017, 9, 682 12 of 13

6. Conclusions
Considering that the conventional ROM overestimates the tensile strength of FRP rebars with
circular cross-section, this study proposed modified ROM coefficients based upon the tensile strength
test results obtained for a series of GFRP rebars with various diameters and hollowness ratios.
According to the type and volume of the materials used, FRP bars each present a different performance,
thus it is not appropriate to extrapolate from this model and apply the results generally. However,
the results of this study are expected to be applicable for use as an index for design and evaluation
purposes by researchers developing GFRP rebars. The conclusions found through this research are
as follows.
(1) For a specific composite that is bar shaped and has a circular cross-section, a more realistic
phenomenon is realized, whereby the tensile force is transferred not through the cross-section
but the side surface of the bar and this resulting uniform stress distribution is included in the
proposed ROM.
(2) The proposed ROM has just one parameter to explain the shear-lag effect of the FRP bar under
tension and the parameter was determined from tensile test results for GFRP rebars with a hollow
section since the fiber content is the main factor for the tensile performance of a FRP outlined in
the conventional ROM.
(3) The proposed ROM provides accuracy within ±5%, whereas the conventional ROM overestimates
the tensile strength of the GFRP reinforcing bars with diameters of 13, 16, and 19 mm by more
than 40%.

Acknowledgments: This project was supported by KICT’s (Korea Institute of Civil Engineering and Building
Technology) basic project named “Harbor Structure’s Performance Improvement Technologies Using FRP
Composites”.
Author Contributions: Young-Jun You and Dong-Woo Seo designed the experiments; Ki-Tae Park guided the
entire work; Jang-Ho Jay Kim and Tae-Hee Lee analyzed the data and corrected the manuscript; and Young-Jun You
wrote the manuscript.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. ISIS Canada. Design Manual 3: Reinforcing Concrete Structures with Fiber Reinforced Polymers; The Canadian
Network of Centers of Excellence on Intelligent Sensing for Innovative Structures: University of Manitoba,
Winnipeg, MB, Canada, 2007.
2. Benmokrane, B.; El-Salakawy, E.; El-Ragaby, A.; Lackey, T. Designing and Testing of Concrete Bridge Decks
Reinforced with Glass FRP Bars. J. Bridge Eng. 2006, 11, 217–229. [CrossRef]
3. VORD. 2014. Available online: http://www.vrod.ca/en/downloads.asp (accessed on 24 April 2014).
4. BPComposite. 2014. Available online: http://www.bpcomposites.com/downloads (accessed on 24
April 2014).
5. You, Y.; Kim, J.; Kim, S.; Park, Y. Methods to enhance the guaranteed tensile strength of GFRP rebar to
900 MPa with general fiber volume fraction. Constr. Build. Mater. 2015, 75, 54–62. [CrossRef]
6. Cusson, R.; Xi, Y. The Behavior of Fiber-Reinforced Polymer Reinforcement in Low Temperature Environmental
Climates; Report No. CDOT-DTD-R-2003-4; Department of Civil, Environmental & Architectural Engineering,
University of Colorado: Boulder, CO, USA, 2002.
7. Lee, D.G.; Jeong, M.Y.; Choi, J.H.; Cheon, S.S.; Chang, S.H.; Oh, J.H. Composite Materials; Hongrung Publishing
Company: Seoul, Korea, 2007; pp. 173–176. (In Korean)
8. Campbell, F.C. Structural Composite Materials; ASM International: Materials Park, OH, USA, 2010;
ISBN 978-1-61503-037-8.
9. Hull, D.; Clyne, T.W. An Introduction to Composite Materials, 2nd ed.; Cambridge University Press: Cambridge,
UK, 2006; pp. 66–67.
10. Lee, C.; Hwang, W. Modified rule of mixtures for prediction of tensile strength of unidirectional
fiber-reinforced composites. J. Mater. Sci. Lett. 1998, 17, 1601–1603. [CrossRef]
Polymers 2017, 9, 682 13 of 13

11. Faza, S.; GangaRao, H. Glass FRP reinforcing bars for concrete. In Fibre-Reinforced Plastic (FRP) Reinforcement
for Concrete Structures: Properties and Applications; Nanni, A., Ed.; Elsevier: Amsterdam, The Netherland,
1993; pp. 167–188.
12. Protasio, F.C.; Nicholas, J.C. Tensile and nondestructive testing of FRP bars. J. Compos. Constr. 1998, 2, 17–27.
13. Shicheng, T. Bond of Glass-Fiber-Reinforced-Plastic Reinforcing Bars to Concrete. Ph.D. Thesis, the University
of Arizona, Tucson, AZ, USA, 1994; p. 53.
14. Ko, F.K.; Somboonsong, W.; Harris, H.G. Fiber architecture based design of ductile composite rebars for
concrete structures. In Proceedings of the 11th International Conference Composite Materials; Scott, M.L., Ed.;
Woodhead Publishing: Cambridge, UK, 1997; p. 4.
15. Canadian Standards Association S806-02. Design and Construction of Building Components with Fiber Reinforced
Polymers; CSA: Toronto, ON, Canada, 2002.
16. ASLAN. 2014. Available online: http://aslanfrp.com/aslan100/Resources/Aslan100a.pdf (accessed on
24 April 2014).
17. Park, K. Development of Enhancing Life Span Technology FRP Waterfront Structures Using FRP Hybrid Bars; Report
No. KICT-2014-162; Korea Institute of Civil Engineering and Building Technology (KICT): Gyeonggi-Do,
Korea, 2014. (In Korean)
18. Kocaoz, S.; Samaranayake, V.A.; Nanni, A. Tensile characterization of glass FRP bars. Compos. Part B 2005,
36, 127–134. [CrossRef]
19. Kim, K. Design and Construction Technology for Concrete Structures Using Advanced Composite Materials; Report
No. KICT-2008-165; Korea Institute of Civil Engineering and Building Technology (KICT): Gyeonggi-Do,
Korea, 2008; p. 214. (In Korean)

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).